Molecular and Cellular Probes 28 (2014) 65e72

Contents lists available at ScienceDirect

Molecular and Cellular Probes

journal homepage: www.elsevier.com/locate/ymcpr

Protein markers for identification of Yersinia pestis and their variation

related to culture
David Wunschel*, Heather Engelmann, Kristin Victry, Brian Clowers, Christina Sorensen,
Nancy Valentine, Christine Mahoney, Thomas Wietsma, Karen Wahl
Pacific Northwest National Laboratory, Richland, WA 99352, USA

a r t i c l e i n f o a b s t r a c t

Article history: The detection of high consequence pathogens, such as Yersinia pestis, is well established in biodefense
Received 11 July 2013 laboratories for bioterror situations. Laboratory protocols are well established using specified culture
Received in revised form media and a growth temperature of 37
C for expression of specific antigens. Direct detection of Y. pestis
27 November 2013
protein markers, without prior culture, depends on their expression. Unfortunately protein expression
Accepted 2 December 2013
Available online 11 December 2013
can be impacted by the culture medium which cannot be predicted ahead of time. Furthermore, higher
biomass yields are obtained at the optimal growth temperature (i.e. 28
Ce30
C) and therefore are more
likely to be used for bulk production. Analysis of Y. pestis grown on several types of media at 30
C
Keywords:
Yersinia pestis
showed that several protein markers were found to be differentially detected in different media. Analysis
Culture environment of the identified proteins against a comprehensive database provided an additional level of organism
Protein expression identification. Peptides corresponding to variable regions of some proteins could separate large groups of
HPLC strains and aid in organism identification. This work illustrates the need to understand variability of
Tandem mass spectrometry protein expression for detection targets. The potential for relating expression changes of known proteins
to specific media factors, even in nutrient rich and chemically complex culture medium, may provide the
opportunity to draw forensic information from protein profiles.
Ó 2013 Elsevier Ltd. All rights reserved.

1. Introduction samples [5]. BHI is most useful for rapid growth of the organism
while cefsulodin-irgasan novobiocin (CIN) agar has been investi-
Confident identification of the potential bioterrorism agent gated as a semi-selective agar for Yersinia isolation [6]. While
Yersinia pestis is critical for a rapid response to an incident. temperature has been well described for its impact on growth rate
Guidelines for culture of suspected samples to yield colonies for and protein synthesis, the impact of different types of complex
microscopy and identification are provided to member labs of the growth medium and their formulations has not. These factors have
laboratory response network (LRN), to demonstrate morphology an important influence on the expression of diagnostic proteins and
and antigen expression characteristics [1]. In particular the rec- culture-based identification.
ommendations call for culture of any suspected isolate at both 28
C Beyond the diagnostic role of the plasmid-borne capsular F1
and 37
C where the former is in a more optimal growth temper- antigen (CaF1), it is important for virulence [7]. Additional viru-
ature range (between 28
C and 30
C) and the later is needed for lence related proteins have been well-studied for their expression
production of the diagnostic F1 capsule antigen. The F1 capsule has with both temperature and calcium concentration. Many of these
also been the focus of clinical diagnostic assays as well [2]. The proteins are plasmid-borne including the “low calcium respon-
recommendations for culture conditions build on previous data se”(lcr) proteins found on pCD1 that are necessary for virulence and
describing the impact of temperature on growth rate and protein are inhibited in their expression with 2.5 mM Ca and culture at
synthesis [4,3]. Complex growth media such as brain heart infusion 26
C [8]. The induction of these proteins was demonstrated at
(BHI), tryptic soy broth (TSB), nutrient broth (NB), sheep blood agar 37
C by Perry and Fetherston [9]. These proteins include the V
(SBA), and Maconkey agar are recommended for culture of clinical antigen and Yersinia outer membrane proteins (yops) including lcr
proteins D-H, R, V, Q and yops BeF, H, and J-M [9]. The optimal
growth temperature for expression of the V antigen (V) has been
* Corresponding author. described at 37
C [10]. Sub millimolar Ca concentrations and
E-mail address: David.Wunschel@pnnl.gov (D. Wunschel). growth at 37
C are thought to mimic the conditions within the

0890-8508/$ e see front matter Ó 2013 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.mcp.2013.12.001
66 D. Wunschel et al. / Molecular and Cellular Probes 28 (2014) 65e72

mammalian host where these protein factors are needed for viru- incubated at 30
for 18 h. Broth cultures were started by inoculating
lence [11,12]. a single colony into 10 mL of appropriate media in a 50 mL glass
In addition to toxins, Fe acquisition and storage has been shown to flask. This starter culture was then sub-cultured 1:100 into 75 mL of
be critical for virulence [13]. Expression of hemin storage proteins the appropriate media in a 250 mL glass flask.
and the siderophore yersiniabactin for acquisition of hemin bound or Two medium recipes were used for organism culture, Brain
free Fe have been identified as specific factors necessary for virulence Heart Infusion (BHI) and Tryptic Soy Broth (TSB). BHI medium was
[9,13]. Loss of the 102 kB pigmentation (pgm) locus coding for these formulated with brain heart extract (17.5 g/L), peptone (10 g/L),
factors has been shown to severely attenuate the organism virulence dextrose (2 g/L), NaCl (5 g/L), and Na2PO4 (2.5 g/L). TSB medium was
[14]. Fe acquisition is an important aspect of bacterial survival in a formulated with tryptone casein digest (15 g/L), soytone protein
variety of environments and so Fe-responsive transcriptional regu- digest (3 g/L), NaCl (5 g/L), and K2HPO4 (2.5 g/L). The BHI and TSB
lators (termed ferric uptake regulators [Fur]) have been found that media were formulated using components purchased separately to
control different mechanisms of Fe acquisition in a variety of path- create “in-house” recipes of each. BHI 1 and BHI 2 were both
ogens [15]. Pieper et al. [16] performed a more extensive study on the formulated using the Heart & Brain Extract of Pork from the Solabia
impact of Fe and temperature on protein expression in Y. pestis. They Group (Pantin, Edex France). BHI1 contained a Bacto Brand Peptone
demonstrated that Fur and RhyB controlled the expression of inde- (BD Biosciences, San Jose, CA USA) while BHI 2 was formulated with
pendent sets of proteins involved in Fe acquisition depending on proteose peptone #3 (Global Bioingredients, Tampa FL, USA). The
whether the organism was grown at 26
C or 37
C. No data was TSB 1 recipe was formulated with a tryptone, “pancreatic digest,
presented on protein expression at 28e30
C. peptone from casein” and a soytone, “peptone from soymeal
The pMT1 and pPCP1 plasmids are specific to Y. pestis among the papain-digested for microbiology” both from Merck KGaA,
Yersinia species and contain the CaF1 in addition to Yersinia murine (Darmstadt, Germany). TSB 2 was formulated with the tryptone,
toxin (ymt) and plasminogen activator protease (Pla) virulence “Bacto Tryptone” (BD Biosciences, San Jose, CA USA) and soytone,
factors. The impact of metals, alone or in conjunction, with changing “SE 50 M enzymatic digest from soy flour” peptone (DMV Inter-
temperature on the expression of these proteins has not been well national bv. Veghel, Netherlands).
defined. Even for those proteins whose expression has been well Cultures of each media were grown in triplicate at 30
C and
studied under defined conditions, the temperatures used have been duplicate at 26
C or 37
C using a rotator incubator at 210 rpm. Late
intended to simulate the host environment [11]. Furthermore, the Fe log phase cultures were harvested at 11 h at OD600 0.8e1.0, sta-
or Ca concentrations studied have been intended to mimic host tionary phase were harvested at 32 h at OD600 1.4e1.8 and again at
conditions or have a definitive impact on protein expression. While 50 h. Cultures were pelleted by centrifugation at 6000 rcf for 10 min
these studies have been vital to our understanding of the interaction and washed in sterile water. The cell supernatant was collected for
between metals, temperature and protein expression, they do not protein precipitation. Final cell suspensions were enumerated on
reflect growth conditions that are optimal for organism production the appropriate solid media. One mL of the supernatant for each
in rich growth media and therefore are not directly relevant for medium was precipitated in a 1.5 mL tube by trichloroacetate (20%
microbial identification and forensics. final volume) and placing the sample at 20
C overnight. The
Y. pestis specific proteins exist for detection but their expression precipitate was pelleted by centrifugation at 16,000 rcf for 5 min
has not been studied in rich media at growth temperatures optimal and 4
C. The pellet was washed in 3  200 mL of ice-cold acetone
for that organism. Beyond species-level identification, differentia- then resuspended in 100 mL of phosphate buffered saline and 0.01%
tion among the Y. pestis isolates has been proposed through iden- SDS.
tification of outer membrane proteins using a proteomic analysis
[17]. However, strain identification and differentiation depends 2.3. Inductively coupled plasma mass spectrometry
on both the presence and sequence variability of the detected
proteins. The presence of essential proteins may be affected by Sterile media samples were prepared for inductively coupled
growth medium, so evaluating the impact of common media is plasma mass spectrometry (ICP-MS). A calibration curve for Be, Na,
essential. In addition to identification, the profile of identified Mg, Al, K, Ca, Ti, V, Cr, Mn, Fe, Co, Ni, Cu, Zn, As, Se, Sr, Mo, Ag, Cd, Sn,
proteins may provide a means to determine attributes of the Sb, Ba, Tl, and Pb was used for elemental composition determina-
original culture medium as a forensic signature. tion. One half mL of each media sample was digested with 4.5 mL of
To that end, we have investigated two versions of two different two percent concentrated nitric acid prior to sample introduction
culture media recipes formulated with components from different into the ICP-MS Model 7500 (Agilent Technologies, Santa Clara CA).
sources for cultivation of a Y. pestis strain. The growth character- Each media sample was analyzed in triplicate with mean and
istics and secreted protein profiles were examined in relation to standard deviation reported for biologically relevant elements in
medium recipe. The ability to discern between Y. pestis sequence Fig. 3.
homologues in the database was also investigated.
2.4. Protein separation and digestion
2. Methods and materials
Samples were prepared for separation by 1-D polyacrylamide
2.1. Chemicals and reagents gel electrophoresis (PAGE) by adding 10 mL of precipitate to 1 mL of
NuPAGE reducing agent and 2.5 mL of NuPAGE loading buffer as
The HPLC grade water, urea, acetonitrile, acetone, ammonium recommended by the manufacturer (Invitrogen, Torrance, CA).
bicarbonate, trichloroacetic acid and formic acid were purchased 10 mL of Precision Plus ProteinÔ Standards (BioRad, Hercules, CA)
from Sigma chemical Co (St. Louis MO). Sequencing grade trypsin was loaded into an outside lane for size estimate. The proteins were
was purchased from Promega (Madison WI). separated on a NuPAGE Bis-Tris 4e12% pre-cast polyacrylamide gel.
The gels were run at 200 V for 35 min and stained with SimplyBlue
2.2. Bacterial strain and growth conditions SafeStain (Invitrogen) for 1 h and destained using deionized water
overnight.
Y. pestis KIM D27 (pgm-) was maintained in glycerol at 80
C Gel bands were excised using an ethanol cleaned razor blade
prior to streaking onto brain heart infusion agar (BHI) plates and and placed into a 1.5 mL microcentrifuge tube. The gel slices were
 D. Wunschel et al. / Molecular and Cellular Probes 28 (2014) 65e72 67

processed for in-gel digestion following Promega’s (Promega Cor- the X!Tandem [18] open source tool against a database comprised
poration, Madison, WI) trypsin protocol (Technical Bulletin for use of the non-redundant proteomes for 36 strains of Yersinia species
of Product V5280 pages 4e5). Proteins were digested in 100 mL of found at the UniProt consortium website (www.uniprot.org). Rep-
4 mg/mL Trypsin Gold in digestion buffer (40 mM NH2HCO3/10% resentatives of 8 Yersinia species were included as follows pestis
acetonitrile) overnight at 37
C. Gel slice extracts were purified (16), Yersinia enterocolitica (7), Yersinia pseudotuberculosis (3), Yer-
using PepCleanÔ C-18 Spin Columns (Thermo Fisher Scientific, sinia intermedia (2), Yersinia bercovieri (2), Yersinia mollaretii (2),
Rockford, IL). Eluted samples were gently dried in a vacuum Yersinia fredriksenii (2) and Yersinia ruckeri (2). Where possible,
evaporator, suspended in 0.1% formic acid, and placed in a vial for protein accession numbers are given for the source strain Y. pestis
liquid chromatography mass spectrometry (LC-MS). KIM D27.
Output from each X!Tandem search was processed directly into
a custom formatted text file compatible with the input re-
2.5. Peptide mass spectrometry and database searching
quirements of the mass spectrometry generating function (MSGF)
provided by Kim et al. [19] using a custom Python script that is
Peptide separation was performed using an Agilent 1200 HPLC
available upon request. Peptide spectrum matches with p-values
system with a 40 cm long 0.15 mm ID fused silica packed with
greater than 1  1010 were excluded from consideration in
Jupiter 5 mm C-18 resin (Phenomenex, Torrence CA). The peptides
accordance with the cutoff suggested by Kim et al. [19] to limit false
were eluted using a gradient separation from 0% to 59% B over
discovery rates to below 0.1%.
60 min where solvent “A” was 5% acetonitrile and 0.1% formic acid
in water and “B” 95% acetonitrile and 0.1% formic acid in water. The
HPLC flow rate of 2 mL per minute was transferred and ionized into 3. Results and discussion
an LTQ Orbitrap system (Thermo electron, Billercia MA) using a
75 mM ID fused silica electrospray emitter (etched to a narrow tip) The growth characteristics of Y. pestis were investigated in
held at þ2.0 kV. Precursor ions were measured in MS mode with a relation to common culture media BHI and TSB. The organism
mass resolution setting of 30,000 and product ions were measured growth in each of two versions of each medium recipe was
in MS/MS for the top five most intense precursor ions in a data- measured using both optical density and colony counts. The growth
dependent fashion using the LTQ analyzer. Dissociation was per- curves using each method are found in Fig. 1. The replicate cultures
formed using 35% normalized collision energy and the default for each version of each medium recipe showed relative similarity
number of microscans. based on plate counts. The growth curves based on plate counts
Each LC-MS dataset was converted into a mascot general format showed most of the cultures entering stationary phase at 12e15 h.
for greater compatibility with database search tools. This procedure The amount of cell mass was 5  108 cfu/mL for the BHI cultures
was used to minimize low intensity contributions while main- while the TSB cultures were closer to 1  108 cfu/mL. By contrast the
taining pertinent spectral features. Data sets were searched using growth curves generated using the optical density measurements

Fig. 1. The number of organisms per mL estimated from plate counts expressed in cfu/mL (A) and optical density expressed in relative absorbance units at 600 nm (B). Two of
versions each medium with two replicate cultures are represented in each plot.
68 D. Wunschel et al. / Molecular and Cellular Probes 28 (2014) 65e72

does not seem to correspond to an overall induction or suppression

of protein bands observed in Fig. 2, there appears to be larger
amounts of Fe in both of the BHI media relative to either TSB recipe.
This appears to correspond to the larger amount of protein
observed from an initial cursory look at the protein content on the
gel for the supernatants from 30
C. It was unknown if specific
proteins present in the supernatant were differentially expressed.
So the gel separation was repeated with a 5 loading and bands at
the 40, 20, 18, 15, and 12 kDa size were excised for digestion and LC-
MS/MS analysis. The bands at 15 and 18 kDa appeared to have
different intensities on the gel between media types, so it was
hypothesized that they may contain different proteins.
Analysis of the resulting peptides derived from the gel bands by
LC-MS/MS corresponded to a complex mixture of proteins. The
combined protein list for all bands and three culture time points for
each medium was compiled. There were 170 and 210 proteins or
subunits identified by at least one peptide through the peptide
digests across all time points for the BHI1 and BHI2 gel slices
Fig. 2. Supernatant protein profile of Y. pestis KIM D27 grown at 30
C on four media respectively. The TSB1 and TSB2 had 98 and 149 proteins repre-
(TSB1,TSB2, BHI1 and BHI2) and taken from different time points. The 11 h time point
sented in the peptide digests by at least one peptide from the five
was only used for the BHI media samples. A protein standard size ladder is provided at
the left. The protein observed at w50 kDa in the BHI media recipes was found to be a sizes of gel slices and three culture time points. In each case, the
component of the porcine brain-heart extract component. 50 h supernatant samples appeared to have the most protein. The
pathogenesis factors and those proteins impacted by Ca and Fe have
been well studied for their expression relative to metal ion and
showed a much larger apparent difference between the culture temperature. Several of the proteins are also pMT1 plasmid borne
media. The BHI1, BHI2, TSB2 showed significantly higher top OD and may be unique to Y. pestis and useful for identification. We
values by about 15 h than TSB-1. The late stationary phase differ- focused on the presence of a small subset of proteins including
ences in OD were greatest at 32 h and longer with BHI-2 having the
highest final OD.
The significant differences in optical density between media
recipes indicated that differences in mass beyond differences in total
cell numbers were present. One hypothesis was that additional
extracellular material was produced in specific medium recipes. As a
result, the supernatants from the cultures were collected and pro-
teins precipitated. A single dimension polyacrylamide gel electro-
phoresis (PAGE) separation was used to make a simple visual
comparison of the proteins based on medium and culture time
(Fig. 2). The relatively simple mixture of proteins within the su-
pernatant was easily compared for differences in total protein and
specific bands among samples from each culture condition.
The amount of protein present in the lanes for BHI1 and BHI2
samples was undetectable at the 11 h time point but abundant at
the 32 h and 50 h time points. These samples appeared to contain
significantly more protein than samples cultivated in either type of
TSB at any time point. The TSB2 samples appeared to have more
protein than the TSB1 samples with detectable bands at 32 h and
more distinct bands at 50 h. The profile of bands in TSB2 appear in a
similar size profile to the 32 h and 50 h lanes for BHI1 and BHI2,
although considerably less concentrated. The highest OD medium
appeared to have the highest protein loading on the gel (BHI2) and
the lowest OD sample the least amount of supernatant protein on
the gel (TSB1). By contrast TSB 2 and BHI1 appear to have similar
growth curves based on cell counts and OD despite the fact that
BHI1 contained more supernatant protein than TSB as judged by
the bands on the gel. The amount of protein in the supernatant
appears to correlate to some degree with the OD achieved by each
culture at 30
C.
The concentration of 20 metals was measured in the sterile
media to determine if the concentration of any metal species cor-
responded with the apparent amount of supernatant protein.
Metals such as sodium and potassium that were separately added
to the media were not included. The relative amount of five metals,
Mg, Ca, Al, Fe and Mn in each medium is presented in Fig. 3. Ca and
Fe in particular have been previously shown to impact protein
expression [11,12,16,20,21]. While the relative concentration of Ca Fig. 3. Comparison of metals measured in each culture medium expressed in mg/L.
 D. Wunschel et al. / Molecular and Cellular Probes 28 (2014) 65e72 69

those related to pathogenesis, Fe acquisition and those present in medium and less distinct band present in the BHI2 sample at 37
C.
the outer membrane (Table 1). These bands were excised and analyzed by LC-MS-MS following
Four proteins suspected of being involved in Fe metabolism (bfn, trypsin digestion and confirmed to contain the F1 capsule antigen
nfuA, Fur, ftnA) were included. These proteins were represented by with 7e24 unique peptides (data not shown).
lower peptide coverage and did not appear dramatically different
between the media. The ferritin (ftn) appeared only in the BHI 3.1. Protein homology and species or strain differentiation
media, but was only represented by a couple peptides. Likewise a
number of outer membrane proteins were identified from the gel The detection of the Y. pestis especific proteins encoded on
bands (Omp C, A, X, yopN and yop E). In some cases no peptides plasmids, such as pMT1 and pPCP1, are important for detection and
were detected for the cells cultured in either one TSB medium type laboratory diagnostics. The detected proteins may also provide
or another. All appear to be present in the BHI media gel slices. The markers for differentiation among strains based on detectable
proteins thought to be vital for pathogenesis including the F1 amino acid variation. The mass spectrometry analysis allows for the
capsule antigen (CaF1), pH 6 antigen (Psa), attachment invasion identified peptides to be compared across multiple Yersinia pro-
locus (Ail), plasminogen activator/coagulase/fibrinolysin (Pla) and teomes in order to determine their taxonomic uniqueness. To that
pesticin protein (Pst) showed differential expression with medium end a previous report stated that peptides derived from the outer
type. In particular the CaF1 and Psa were represented by a large membrane protein fraction of Y. pestis were “strain-unique” and
number of unique peptides in the BHI media, while they were ab- allowed strain-level identification [17]. Our approach differs from
sent in either TSB medium. In general the TSB1 medium poorly the previous report because we have compared our results to a
produced any of the pathogenesis proteins. The two TSB media broader database of Yersinia species protein sequences and targeted
appeared different for some proteins where Ail and Pst were seen specific proteins as markers to distinguish between strains present
with significantly more peptides in the TSB2 medium but near in the database. In practice we make comparisons at the protein
absent in TSB1. level by searching for the database protein sequences containing
The proteins detected in the culture supernatants at 30
C were the largest number of peptides identified within an experimental
the focus of the media comparison, however the expression of dataset relative to the homologous sequences from other organ-
specific pathogenesis factors have previously been described to isms. These proteins, and therefore their source organisms, are
occur at 37
C and not at temperatures below 32
C. The organism more likely to be the true source of the peptides. In this way, the
was also grown on the same medium at the typically studied likely strain or species can be determined for an unknown sample
growth temperatures 26
C and 37
C to the 24 h and 43 h time using proteomic data and searching for the variable number of
points as the respective stationary phase for each culture medium. “matching” peptides between homologues.
The supernatants were harvested, precipitated and separated by 1- A subset of the proteins used by Jabbour et al. [17] were detected
D PAGE for the cultures. The TSB2 and BHI2 supernatants for the in the culture supernatants and included Omp C (Q0WHI4), CaF1
26
C and 37
C cultures are shown in Fig. 4. Very few of the same (D1U2C7), and Pla (D1U2P5). Additionally the proteins Omp A
bands are observed at 26
C for either medium relative to the 30
C (D1U1B0) and Ail (D1TU90) were proteins identified with high
cultures. A distinct band at 18 kDa was observed for the TSB2 confidence using multiple peptides in at least 3 of the media. These

Table 1
Selected proteins identified across time points in each type of medium at 30
C. The largest gel band the protein was detected in is provided along with the UniProt accession
number and maximum number of unique peptides identified for that protein in each medium recipe across the three time points. The protein groups involved in iron
metabolism, pathogenesis and present in the outer membrane are differentially shaded.

Gel Band Uniprot ID Protein Description Media

kDa BHI1 BHI2 TSB1 TSB2
18 Q7CFU0 Bacterioferritin (Bfn) 5 6 2 7

20 Q8ZJI0 Fe/S biogenesis protein (nfuA) 4 4 3 3

Ferric uptake regulation protein
18 P33086 (Fur) 2 2 0 2
18 Q0WG08 Ferritin (ftn) 1 2 0 0

20 P17811 Plasminogen activator (Pla) 2 10 2 6

18 Q0WCZ9 Attachment invasion locus (Ail) 3 17 2 22

18 P26948 F1 capsule antigen (CaF1) 13 18 0 0
40 Q57159 Pesticin (Pst) 6 11 1 23
15 P31522 pH 6 antigen (Psa) 21 6 0 0

12 P69957 Type III secretion protein (LcrG) 0 0 1 1

Virulence-associated V antigen
20 P0C7U7 (LcrV) 0 0 0 2
Low calcium response protein H
18 P21207 (LcrH) 0 1 0 2
40 Q7CHA0 Outer membrane protein C/1b 9 15 4 3
40 Q8ZG77 Outer membrane protein A 9 9 9 11
18 Q8D0S1 Outer membrane protein X 7 5 0 6
20 P68640 Outer membrane protein yopN 2 1 2 5
20 P31493 Outer membrane protein yopE 5 2 0 9
70 D. Wunschel et al. / Molecular and Cellular Probes 28 (2014) 65e72

Y. pestis isolates and distinguish KIM D27 from a small cluster of

strains in the database.
Pla was found to have 81 protein homologues from Y. pestis
strains that contained 100% sequence identity while 2 strains had
homologues containing a single amino acid substitution. The next
closest homologous sequences were from Klebsiella and Escherichia
species which had Pla or Pla-like (coagulase/fibrinolysin) proteins
with 80% sequence identity. Within the 10 peptides identified for
Pla in the BHI2 sample, four of them correspond to the residues 21e
59 which distinguish between Pla from Y. pestis strains and Pla from
more distantly related Klebsiella species. The high degree of
sequence conservation uncovered through BLAST sequence com-
parisons for this protein make it unsuitable for differentiation
among all known strains of Y. pestis.
The 15 peptides identified for Omp C were compared to the
UniProt database were found to be highly conserved among the
strains of Y. pestis and Y. pseudotuberculosis. A whole protein BLAST
comparison of the accession number Q0WHI4 found 92 entries
among the strains of the two species with 100% sequence for Omp C.
Several strains from Yersinia species such as intermedia, bercovieri,
mollaretii had sequence identities of w84% while enterocolitica
strains had sequence identity of less than 80%. Similarly, Omp A was
found to be highly conserved with 89 homologues from Y. pestis and
Yersinia pseudouberculosis in the database having 100% sequence
identity. A number of other Yersinia species had sequence identity at
the 88%e92% level. Consistently identified peptides corresponded to
amino acid residues 140e161, 197e221 and 226e234 distinguished
Y. pestis and Y. pseudotuberculosis species from other Yersinia species.
However, based on the BLAST sequence comparison results, these
two proteins provide no utility in differentiating among the strains
of Y. pestis and Y. pseudotuberculosis.
Two other pathogenesis related proteins, Psa and pesticin pro-
Fig. 4. Comparison of supernatant proteins at 26
C and 37
C. tein (Pst), were identified from the gel bands that were investigated
for their ability to discriminate among the Yersinia using the
identified peptides. Performing a blast search for Pst (D1U2P1)
were examined for their ability to differentiate between Yersinia resulted in 83 strains of Y. pestis matching with 100% sequence
species and potentially Y. pestis strains. CaF1 provides species identity to one another, while four strains have a single amino acid
specificity because it present on the pMT1 plasmid which is found substitution, including the strain examined in this study (KIM D27).
in only isolates of Y. pestis. No peptide could be identified containing the amino acid substi-
Although the CaF1 protein is specific to Y. pestis as a plasmid- tution that differentiates between this strain and the bulk of
borne marker, a BLAST search of CaF1 (D1U2C7) against the Uni- Y. pestis strains. Similarly, the Psa protein (D1TPJ6) had 89 homol-
Prot database revealed that it is also highly conserved within that ogous proteins from both Y. pestis and Y. pseudotuberculosis with
species’ isolates. At least 27 homologues were matched with 100% 100% sequence identity within the database. Only a few strains of
sequence conservation while 57 partial sequences contain a single either species contained a single amino acid substitution at position
amino acid substitution at position 129. There are three full se- 151 and a distinguishing peptide covering residues 148e158 was
quences with a single serine to alanine amino acid substitution at identified in both BHI1 and BHI2 supernatant samples. The high
position 48. The peptides corresponding to amino acid residues 46e degree of sequence conservation for these proteins makes them
73 and 123e145 were both consistently detected in the BHI1 and BHI generally unsuitable for differentiation among almost all sequences
2 supernatants. Furthermore, two peptides were identified covering from strains of Y. pestis.
positions 46e73, one with the serine and the other an alanine at
position 48 making the KIM D27 strain used in this study indistin- 4. Conclusions
guishable from the majority of other strains using this protein.
Unlike CaF1, the Ail is a pathogenesis factor with homologues The growth conditions used to cultivate a pathogen can
present in the other major human Yersinia pathogenic species dramatically impact protein expression. When protein targets are
Y. enterocolitica and Y. pseudotuberculosis. A BLAST search using the used for identification of pathogens, the impact of culture condi-
KIM D27 sequence (D1TU90) showed that sequence diversity exists tions on protein expression needs to be considered. The protein
where all the Y. enterocolitica Ail homologues, and all but a single expression has been well studied for Y. pestis under either defined
Y. pseudotuberculosis strain, can be distinguished from the 81 culture conditions to better understand the pathogen biology, or in
Y. pestis strains. Using the identified peptides, and particularly the clinical diagnostics using a defined protocol specifying growth at
peptide covering residue positions 132e142, the Y. pestis strains can 37
C. However, the impact of complex culture media recipes and
be distinguished from other Yersinia species. Up to 13 strains of more optimal growth temperatures (28e30
C) on protein
Y. pestis were distinct from the larger group of Y. pestis Ail homo- expression have not been as thoroughly studied. For direct detec-
logues by substitutions within that region and also the peptide tion (without prior culture) in forensic and biodefense situations,
covering amino acid residues 145e177. The result is that this pro- these are the most relevant production conditions to examine.
tein marker can provide some differentiation among the known Several of the proteins, including CaF1, Psa and Pla, have been
 D. Wunschel et al. / Molecular and Cellular Probes 28 (2014) 65e72 71

previously described as exclusively expressed at 37
C [23]. That different Yersinia species and to a limited extent different between
does not appear to be strictly true based on results presented here, groups of strains. However the data must be compared against a
although the expression of the proteins at 30
C does appear to be comprehensive sequence database to understand the taxonomic
dependent on the growth medium used. value of individual proteins. A subject of future study is a more
In order to better explore the breadth of protein expression, two comprehensive proteomic analysis of the cellular proteome to
common recipes were formulated with different sources of the examine a larger set of protein markers.
basic constituents. Significant differences appeared in the growth In addition to organism identification, evaluating protein
characteristics depending on both media recipe and formulation. expression based on specific media factors may be useful as a
These differences were reflected in the amount and type of protein forensic tool to aid in determining the type of media used to pro-
observed in the culture supernatant at various time points. A duce a sample. Genetic analysis has been used in the past for
number of gel bands were apparent and five sizes were selected for exquisitely accurate typing of Y. pestis isolates and provides a means
in-depth proteomic analysis. From 98 to 210 proteins, depending on to place isolates on a phylogeographic map [26]. While this type of
medium type, were identified by at least one peptide across the data is a critically important part of microbial forensics, it provides
three time points, temperatures and five gel slices. From these no information on the immediate culture history. Detection and
larger lists of proteins, seventeen proteins were used for compari- identification of peptides derived from the culture media has also
son among the four medium types based on their roles in Fe been demonstrated for Y. pestis cell mass and would be another
metabolism, pathogenesis or outer membrane location. Five of the valuable piece of information to determine the likely type of me-
pathogenesis-related proteins (described below) were found to dium used for culture [27]. However analysis of residual medium
differ in the number of unique peptides observed between medium peptides may not provide any information on additional culture
recipe samples. variables such as time or metal content.
The media was analyzed for differences in metal content in or- Differences in the expression of specific pathogenesis-related
der to investigate specific factors influencing growth and protein proteins was observed in relation to the medium used and illus-
expression. Previous studies had focused on Fe and Ca and so they trates the possible role of tying specific protein expression changes
were among the metals analyzed. The expression of specific pro- to culture variables. Even though the concentration of calcium has
teins, particularly F1 capsule antigen and pH 6 antigens were been well known to impact the expression of virulence related
preferentially expressed in the Fe rich BHI media recipes at 30
C. proteins at 37
C, the concentration of Fe appeared to impact
Some proteins related to Fe metabolism were also detected in BHI expression of CaF1 and Psa at 30
C. Further study is required to
media. The BHI recipes contained relatively higher amounts of Fe firmly establish specific and quantitative relationships between
than the TSB recipes, although the concentration of Fe in any of the expression of known proteins (and pathways) and variables in the
four media was well above the “Fe limited” conditions and below culturing process, such as medium composition. To further expand
the “Fe enriched” media investigated by Pieper et al. [16]. Other on this hypothesis, a quantitative analysis of the cellular proteome
“low Ca response” proteins were represented by peptides detected of Y. pestis KIM D27 cultured on the same medium samples is a
only in the TSB media. Several of the outer membrane proteins subject of future study to relate a larger set of protein markers to
were observed with similar peptide coverage in all four media. the time of culture and concentration of metals in these media
Observation of those proteins is consistent with previous reports recipes.
examining the impact of Ca on protein expression [11].
Three of these proteins, CaF1, Omp C and Pla were also described
previously to be “strain-unique” and therefore useful in specifically Acknowledgments
differentiating Y. pestis CO92 strain from other Y. pestis strains
within a database [17]. CaF1 and Pla provide species-level identi- Funding for this work was provided in part through contract
fication because their plasmid distribution is limited to Y. pestis HSHQPM-10-X-00077/4 to Pacific Northwest National Laboratory
strains. Unfortunately, both proteins are highly conserved and a by the Department of Homeland Security Science and Technology
BLAST analysis demonstrated that only a few single amino acid Directorate. Battelle Memorial Institute operates Pacific Northwest
substitutions are present in CaF1 that could possibly differentiate National Laboratory for the U.S. DOE under Contract DE-AC06-
between three large groups of strains that have protein homo- 76RLO.
logues present within the UniProt database. This finding corre-
sponds with a study on the DNA recovered from victims of the Black
Death, where the amplified DNA for these two loci was consistent References
to modern isolates [24]. Furthermore, those plasmid-borne char-
acteristics can be lost during culture and may hamper rapid iden- [1] Sharp SE, Saubolle MA. Sentinel laboratory guidelines for suspected agents of
bioterrorism: Yersinia pestis. Washington D.C: American Society for Microbi-
tification [25]. Even more challenging were the Omp A and C ology; 2009. http://www.asm.org/images/pdf/Clinical/ypestis12-11-09.pdf.
proteins which have homologues with complete sequence con- [2] Sha J, Endsley JJ, Kirtley ML, Foltz SM, Huante MB, Erova TE, et al. Charac-
servation across all available Y. pestis and Y. pseudotuberculosis terization of an F1 deletion mutant of Yersinia pestis CO92, pathogenic role of
F1 antigen in bubonic and pneumonic, and evaluation of sensitivity and
strains. These findings show that strain-specific identification is not specificity of F1 antigen capture-based plague dipsticks. J Clin Microbiol
possible using the identified proteins, including several previously 2011;49(5):1708.
proposed as markers for strain identification [17]. A more [3] Higuchi K, Carlin CE. Studies on the nutrition and physiology of Pasturella
pestis. J Bacteriol 1957;73:122.
comprehensive analysis of the proteome may reveal variable pep- [4] Fukui GM, Ogg JE, Wessman GE, Surgalla MJ. Studies on the relation of cultural
tides useful in strain identification; however any proposed “strain conditions and virulence of Pasturella pesits. J Bacteriol 1957;74(6):714e7.
specific” peptides need to also be confirmed by a in silico compar- [5] Lindsey KQ, Morse SA. Basic laboratory protocols for the presumptive iden-
tification of Yersinia pestis. Centers for Disease Control and Prevention; April
ison to the available Y. pestis and Y. pseudotuberculosis genomes.
18, 2001.
This work demonstrates the need to evaluate the impact of rich [6] Ber R, Mamroud E, Aftalion M, Tidhar A, Gur D, Flashner Y, et al. Development
culture media at optimal growth temperatures on expression of of an improved selective agar medium for isolation of Yersinia pestis. Appl
targeted protein markers. Several proteins have inherent utility in Environ Microbiol 2003;69(10):5787e92.
[7] Sebbane F, Jarrett C, Gardner D, Long D, Hinnebusch BJ. The Yersinia pestis
organism detection if expressed. Proteomic analysis can augment caf1M1A1 fimbrial capsule operon promotes transmission by flea bite in a
identification by distinguishing between sources of proteins from mouse model of bubonic plague. Infect Immun 2009;77(3):1222e9.
72 D. Wunschel et al. / Molecular and Cellular Probes 28 (2014) 65e72

[8] Zahorchak RJ, Brubaker RR. Effect of exogenous nucleotides on Ca20 depen- [17] Jabbour R, Wade MM, Deshpande SV, Stanford MF, Wick CH, Zulich AW, et al.
dence and V Antigen synthesis in Yersinia pestis. Infect Immun 1982;38(3): Identification of Yersinia pestis and Escherichia coli strains by whole cell and
953e9. outer membrane protein extracts with mass spectrometry-based proteomics.
[9] Perry, Fetherston. Yersinia pestis e etiologic agent of plague. Clin Microbiol J Prot Res 2010;9:3647e55.
Rev 1997;10(1):35e66. [18] Craig R, Beavis RC. Bioinformatics 2004;20(9):1466e7.
[10] Fowler JM, Brubaker RR. Physiological basis of the low calcium response in [19] Kim S, Gupta N, Pevzner PA. J Prot Res 2008;7(8):3354e63.
Yersinia pestis. Infect Immun 1994;62(12):5234e41. [20] Lucier TS, Fetherston JD, Brubaker RR, Perry RD. Iron uptake and iron-
[11] Valentina A, Feodorova, Golova AB. Antigenic and phenotypic modifications of repressible polypeptides in Yersinia pestis. Infect Immun 1996;64(8):3023e31.
Yersinia pestis under calcium and glucose concentrations simulating the [21] Perry, Featherston. Yersiniabactin iron uptake: mechanisms and role in Yer-
mammalian bloodstream environment. J Med Microbiol 2005;54:435e41. sinia pestis pathogenesis. Microbes Infect 2011;13:808e17.
[12] Hixson KK, Adkins JN, Baker SE, Moore RJ, Chromy BA, Smith RD, et al. [23] Schriefer ME, Petersen JM. Yersinia. Manual of clinical microbiology. 10th ed.
Biomarker candidate identification in Yersinia pestis using organism-wide In: Versalovic J, Carroll KC, Funke G, Jorgensen JH, Landry ML, Warnock DW,
Semiquantitative proteomics. J Prot Res 2006;5(11):3008e17. editors. vol. 1. p. 627e38.
[13] Perry RD, Lucier TS, Sikkema DJ, Brubaker RR. Storage reservoirs of hemin and [24] Haensch S, Bianucci R, Signoli M, Rajerison M, Schultz M. Distinct clones of
inorganic iron in Yersinia pestis. Infect Immun 1993;61(1):32. Yersinia pestis caused the black death. PLoS Pathog 2010;6(10):e1001134.
[14] Buchrieser C, Rusniok C, Frangeul L, Couve E, Billault A, Kunst K, et al. The 102- [25] Lowell JL, Zhansarina A, Yockey B, Meka-Mechenko T, Stybayeva G, Atshabar B,
Kilobase pgm locus of Yersinia pestis: sequence analysis and comparison of et al. Phenotypic and molecular characterizations of Yersinia pestis isolates from
selected regions among different Yersinia pestis and Yersinia pseudotubercu- Kazakhstan and adjacent regions. Microbiology 2007;153(1):169e77.
losis strains. Infect Immun 1999;67(9):4851e61. [26] Keim PS, Wagner DM. Humans and evolutionary and ecological forces shaped
[15] Carpenter BM, Whitmire JM, Merrell DS. This is not your mother’s repressor: the phylogeography of recently emerged diseases. Nature Rev Micro
the complex role of Fur in pathogenesis. Infect Immun 2009;77(7):2590e601. 2009;8(13):813e21.
[16] Pieper R, Huang ST, Parmar PP, Clark DJ, Alami H, Fleischmann RD, et al. [27] Clowers BH, Wunschel DS, Kreuzer HW, Engelmann HE, Valentine N, Wahl KL.
Proteomic analysis of iron acquisition, metabolic and regulatory responses of Characterization of residual medium peptides from Yersinia pestis cultures.
Yersinia pestis to iron starvation. BMC Microbiol 2010;10:30. Anal Chem 2012;85(8):3933e9.